An Innovative Solution Based on TSCA-ViT for Osteosarcoma Diagnosis in Resource-Limited Settings
Abstract
:1. Introduction
- (1)
- We deployed a novel directed filtering algorithm to improve the quality of pathological image data, which is often affected by various factors like device noise and accidental errors in the sample preparation process. By using the grayscale image of the original picture as guidance, our approach effectively eradicates noise, thereby preserving the image’s texture and offering excellent edge detection. This enhancement in data quality, in turn, significantly boosts the performance of our deep learning model.
- (2)
- Our work introduces an innovative Transformer framework equipped with an efficient twin attention mechanism for sophisticated modeling and segmentation. Utilizing patch-embedding modules, our network attains overlap** patch labels, which are then encoded through an encoder module to gain hierarchical, multiscale representations. Further, we optimized the attention mechanism for efficiency, substantially reducing computational complexity while ensuring high representativeness.
- (3)
- We devised skip connection paths integrated with cross-attention modules to furnish each decoder with spatial information. Coupled with the efficient attention mechanism, this strategy strengthens the model’s localization ability, thereby improving the final model’s precision while maintaining model reusability.
- (4)
- We experimented with a dataset of 1000 pathological images from the Second People’s Hospital of Huaihua. Our experimental findings showcase the advantage of our method in comparison to other convolutional and non-convolutional segmentation networks in segmenting the nuclei of osseous neoplasm cells in pathological sections.
2. Related Work
3. System Model
3.1. Image Denoising
Algorithm 1: Guided Filtering algorithm |
Input: filtering input image , guidance image , radius , regularization Output: filtering output End for; End for |
3.2. Image Analysis and Prediction
4. Simulation Analysis
4.1. Experiment Details
4.2. Assessment Measures
4.3. Training Strategy
4.4. Results
4.5. Discussion
5. Conclusions
Author Contributions
Funding
Institutional Review Board Statement
Informed Consent Statement
Data Availability Statement
Conflicts of Interest
References
- Liu, F.; Zhu, J.; Lv, B.; Yang, L.; Sun, W.; Dai, Z.; Gou, F.; Wu, J. Auxiliary Segmentation Method of Osteosarcoma MRI Image Based on Transformer and U-Net. Comput. Intell. Neurosci. 2022, 2022, 9990092. [Google Scholar] [CrossRef]
- Bray, F.; Laversanne, M.; Weiderpass, E.; Soerjomataram, I. The Ever-Increasing Importance of Cancer as a Leading Cause of Premature Death Worldwide. Cancer 2021, 127, 3029–3030. [Google Scholar] [CrossRef] [PubMed]
- Amiryaghoubi, N.; Fathi, M.; Barar, J.; Omidian, H.; Omidi, Y. Advanced Nanoscale Drug Delivery Systems for Bone Cancer Therapy. Biochim. Biophys. Acta (BBA)-Mol. Basis Dis. 2023, 1869, 166739. [Google Scholar] [CrossRef] [PubMed]
- Yuan, T.; Zeng, J. A Medically Assisted Model for Precise Segmentation of Osteosarcoma Nuclei on Pathological Images. IEEE J. Biomed. Health Inform. 2023, 27, 3982–3993. [Google Scholar] [CrossRef]
- He, K.; Tang, H. Global Field of View-Based Pixel-Level Recognition Method for Medical Images. J. Intell. Fuzzy Syst. 2023, 45, 4009–4021. [Google Scholar] [CrossRef]
- Zhou, Z.; Tan, Y. A Cascaded Multi-Stage Framework for Automatic Detection and Segmentation of Pulmonary Nodules in Develo** Countries. IEEE J. Biomed. Health Inform. 2022, 26, 5619–5630. [Google Scholar] [CrossRef]
- Ouyang, T.; Yang, S.; Dai, Z. Rethinking U-Net from an Attention Perspective with Transformers for Osteosarcoma MRI Image Segmentation. Comput. Intell. Neurosci. 2022, 2022, 7973404. [Google Scholar] [CrossRef] [PubMed]
- Qu, Y.; Li, X.; Yan, Z.; Zhao, L.; Zhang, L.; Liu, C.; ** Countries. Comput. Math. Methods Med. 2022, 2022, 7703583. [Google Scholar] [CrossRef]
- Shen, Y.; Dai, Z. Osteosarcoma MRI Image-Assisted Segmentation System Base on Guided Aggregated Bilateral Network. Mathematics 2022, 10, 1090. [Google Scholar] [CrossRef]
- Ling, Z.; Yang, S.; Dai, Z. Intelligent Assistant Diagnosis System of Osteosarcoma MRI Image Based on Transformer and Convolution in Develo** Countries. IEEE J. Biomed. Health Inform. 2022, 26, 5563–5574. [Google Scholar] [CrossRef] [PubMed]
- Gou, F.; Wu, J. Novel data transmission technology based on complex IoT system in opportunistic social networks. Peer-to-Peer Netw. Appl. 2023, 16, 571–588. [Google Scholar] [CrossRef]
- Tian, X.; Jia, W. Optimal Matching Method Based on Rare Plants in Opportunistic Social Networks. J. Comput. Sci. 2022, 64, 101875. [Google Scholar] [CrossRef]
- Wu, J.; ** Countries. Comput. Intell. Neurosci. 2022, 2022, 7285600. [Google Scholar] [CrossRef]
- Ahmed, I.; Jeon, G.; Piccialli, F. From Artificial Intelligence to Explainable Artificial Intelligence in Industry 4.0: A Survey on What, How, and Where. IEEE Trans. Ind. Inform. 2022, 18, 5031–5042. [Google Scholar] [CrossRef]
- **ao, P.; Huang, H.; Zhou, Z.; Dai, Z. An Artificial Intelligence Multiprocessing Scheme for the Diagnosis of Osteosarcoma MRI Images. IEEE J. Biomed. Health Inform. 2022, 26, 4656–4667. [Google Scholar]
- Yu, L. Data Transmission Scheme Based on Node Model Training and Time Division Multiple Access with IoT in Opportunistic Social Networks. Peer-to-Peer Netw. Appl. 2022, 15, 2719–2743. [Google Scholar]
- Dong, C.; Loy, C.C.; He, K.; Tang, X. Learning a Deep Convolutional Network for Image Super-Resolution. In Computer Vision—ECCV 2014; Fleet, D., Pajdla, T., Schiele, B., Tuytelaars, T., Eds.; Springer International Publishing: Cham, Switzerland, 2014; pp. 184–199. [Google Scholar]
- Ronneberger, O.; Fischer, P.; Brox, T. U-net: Convolutional networks for biomedical image segmentation. In Medical Image Computing and Computer-Assisted Intervention—MICCAI 2015: 18th International Conference, Munich, Germany, 5–9 October 2015, Proceedings, Part III 18; Springer International Publishing: Cham, Switzerland, 2015. [Google Scholar]
- Dai, J.; Li, Y.; He, K.; Sun, J. R-FCN: Object Detection via Region-Based Fully Convolutional Networks. In Advances in Neural Information Processing Systems; Curran Associates, Inc.: Red Hook, NY, USA, 2016; Volume 29. [Google Scholar]
- Liang, Q.-K. Feasibility of deep learning image-based segmentation algorithm in pathological section of gastric cancer. Acad. J. Second. Mil. Med. Univ. 2018, 39, 903–908. [Google Scholar]
- Huang, J.; Gou, F.; Wu, J. An effective data communication community establishment scheme in opportunistic networks. IET Commun. 2023, 17, 1354–1367. [Google Scholar] [CrossRef]
- Dosovitskiy, A.; Beyer, L.; Kolesnikov, A.; Weissenborn, D.; Zhai, X.; Unterthiner, T.; Dehghani, M.; Minderer, M.; Heigold, G.; Gelly, S.; et al. An Image Is Worth 16x16 Words: Transformers for Image Recognition at Scale. ar**v 2021, ar**v:2010.11929. [Google Scholar]
- You, C.; Zhao, R.; Liu, F.; Dong, S.; Chinchali, S.; Topcu, U.; Staib, L.; Duncan, J.S. Class-Aware Adversarial Transformers for Medical Image Segmentation. ar**v 2022, ar**v:2201.10737. [Google Scholar]
- Chen, J.; Lu, Y.; Yu, Q.; Luo, X.; Adeli, E.; Wang, Y.; Lu, L.; Yuille, A.L.; Zhou, Y. TransUNet: Transformers Make Strong Encoders for Medical Image Segmentation. ar**v 2021, ar**v:2102.04306. [Google Scholar]
- Ding, M.; **ao, B.; Codella, N.; Luo, P.; Wang, J.; Yuan, L. DaViT: Dual Attention Vision Transformers. In Computer Vision—ECCV 2022; Avidan, S., Brostow, G., Cissé, M., Farinella, G.M., Hassner, T., Eds.; Lecture Notes in Computer Science; Springer Nature Switzerland: Cham, Switzerland, 2022; Volume 13684, pp. 74–92. ISBN 978-3-031-20052-6. [Google Scholar]
- Huang, X.; Deng, Z.; Li, D.; Yuan, X. MISSFormer: An Effective Medical Image Segmentation Transformer. ar**v 2021, ar**v:2109.07162. [Google Scholar] [CrossRef] [PubMed]
- Cao, H.; Wang, Y.; Chen, J.; Jiang, D.; Zhang, X.; Tian, Q.; Wang, M. Swin-Unet: Unet-like Pure Transformer for Medical Image Segmentation. ar**v 2021, ar**v:2105.05537. [Google Scholar]
- Azad, R.; Arimond, R.; Aghdam, E.K.; Kazerouni, A.; Merhof, D. DAE-Former: Dual Attention-Guided Efficient Transformer for Medical Image Segmentation. ar**v 2023, ar**v:2212.13504. [Google Scholar]
- Eppert, K.; Wunder, J.S.; Aneliunas, V.; Kandel, R.; Andrulis, I.L. von Willebrand factor expression in osteosarcoma metastasis. Mod. Pathol. 2005, 18, 388–397. [Google Scholar] [CrossRef] [PubMed]
- Prediction of Anticancer Peptides Based on an Ensemble Model of Deep Learning and Machine Learning Using Ordinal Positional Encoding. Brief. Bioinform. 2023, 24, bbac630. Available online: https://academic.oup.com/bib/article/24/1/bbac630/6987656?login=true#no-access-message#no-access-message (accessed on 16 September 2023).
- Deng, H.; Eftekhari, Z.; Carlin, C.; Veerapong, J.; Fournier, K.F.; Johnston, F.M.; Dineen, S.P.; Powers, B.D.; Hendrix, R.; Lambert, L.A.; et al. Development and validation of an explainable machine learning model for major complications after cytoreductive surgery. JAMA Netw. Open 2022, 5, e2212930. [Google Scholar] [CrossRef]
- Huang, H.; Lin, L.; Tong, R.; Hu, H.; Zhang, Q.; Iwamoto, Y.; Han, X.; Chen, Y.-W.; Wu, J. UNet 3+: A Full-Scale Connected UNet for Medical Image Segmentation. In Proceedings of the ICASSP 2020—2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Barcelona, Spain, 4–8 May 2020; pp. 1055–1059. [Google Scholar]
- Takaya, E.; Takeichi, Y.; Ozaki, M.; Kurihara, S. Sequential Semi-Supervised Segmentation for Serial Electron Microscopy Image with Small Number of Labels. J. Neurosci. Methods 2021, 351, 109066. [Google Scholar] [CrossRef]
- Dong, X.; Bao, J.; Chen, D.; Zhang, W.; Yu, N.; Yuan, L.; Chen, D.; Guo, B. CSWin Transformer: A General Vision Transformer Backbone with Cross-Shaped Windows. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), New Orleans, LA, USA, 18–24 June 2022; pp. 12124–12134. [Google Scholar]
- Liu, F.; Gou, F.; Wu, J. An Attention-Preserving Network-Based Method for Assisted Segmentation of Osteosarcoma MRI Images. Mathematics 2022, 10, 1665. [Google Scholar] [CrossRef]
Symbol | Paraphrase |
---|---|
Efficient attention | |
Global context vectors | |
Query | |
One parameter of | |
after linear conversion | |
Normalization functions for the keys | |
Normalization functions for the values | |
is the embedding dimension | |
The embedding dimension of queries | |
The embedding dimension of keys | |
The embedding dimension of values | |
Output of the decoder layer | |
Output of the encoder layer | |
after linear conversion |
Model | Ac | Pr | Re | DSC | F1 | IoU | Params | FOLPs | AUC |
---|---|---|---|---|---|---|---|---|---|
Attention-Unet | 0.945 | 0.699 | 0.721 | 0.801 | 0.710 | 0.581 | 34.88 M | 533.08 G | 0.897 |
CENet | 0.933 | 0.646 | 0.868 | 0.696 | 0.691 | 0.556 | 29.53 M | 71.2 G | 0.893 |
CSwin-transfomer | 0.806 | 0.802 | 0.806 | 0.788 | 0.793 | 0.710 | 52.15 M | 230.69 G | 0.886 |
R2U-Net | 0.938 | 0.726 | 0.682 | 0.584 | 0.703 | 0.446 | 35.84 M | 389.17 G | 0.907 |
SegNet | 0.954 | 0.780 | 0.769 | 0.736 | 0.774 | 0.598 | 29.44 M | 3759.14 G | 0.887 |
SERT | 0.954 | 0.712 | 0.741 | 0.702 | 0.726 | 0.551 | 86.21 M | 387.9 G | 0.945 |
Swin-Unet | 0.965 | 0.614 | 0.913 | 0.721 | 0.734 | 0.571 | 27.17 M | 11.74 G | 0.941 |
UNet++ | 0.951 | 0.713 | 0.851 | 0.742 | 0.776 | 0.605 | 9.16 M | 277.26 G | 0.910 |
U-Net | 0.955 | 0.740 | 0.803 | 0.734 | 0.770 | 0.592 | 7.77 M | 110.02 G | 0.878 |
Our (TSCA-ViT) | 0.975 | 0.789 | 0.896 | 0.853 | 0.832 | 0.616 | 49.15 M | 205.11 G | 0.952 |
Our (denoise+TSCA-ViT) | 0.977 | 0.803 | 0.893 | 0.855 | 0.834 | 0.619 | 49.15 M | 210.23 G | 0.952 |
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
© 2023 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).
Share and Cite
He, Z.; Liu, J.; Gou, F.; Wu, J. An Innovative Solution Based on TSCA-ViT for Osteosarcoma Diagnosis in Resource-Limited Settings. Biomedicines 2023, 11, 2740. https://doi.org/10.3390/biomedicines11102740
He Z, Liu J, Gou F, Wu J. An Innovative Solution Based on TSCA-ViT for Osteosarcoma Diagnosis in Resource-Limited Settings. Biomedicines. 2023; 11(10):2740. https://doi.org/10.3390/biomedicines11102740
Chicago/Turabian StyleHe, Zengxiao, Jun Liu, Fangfang Gou, and Jia Wu. 2023. "An Innovative Solution Based on TSCA-ViT for Osteosarcoma Diagnosis in Resource-Limited Settings" Biomedicines 11, no. 10: 2740. https://doi.org/10.3390/biomedicines11102740